The present invention relates generally to the field of program execution management and more particularly to workload scheduling in a heterogeneous environment.
Many modern computers utilize specialized components such as graphical programming units (GPUs) or field-programmable gate arrays (FPGAs) in addition to central processing units (CPUs) in order to perform a specific type of computation more quickly or efficiently. Various data storage devices such as hard disk drives (HDDs), solid state drives (SSDs), or flash storage can also be utilized to improve a data transfer rate or latency.